Lexical Borrowing in the Speech of First-Generation Hungarian Immigrants in Australia

This article reports findings of a sociolinguistic project which investigated language contact phenomena in the speech of first-generation Hungarian Australians living in Sydney. The research aimed to identify and analyze English lexical items borrowed into the spoken Hungarian of first-generation Hungarian–English bilinguals. This research had a mixed methods approach including a quantitative element (count of lexical manifestations by categories such as part of speech) and a qualitative element in which the various lexical manifestations have been subjected to a linguistic analysis. The Hungarian National Corpus was used as a reference guide to determine the status of these phenomena in the lexicon of Standard Hungarian. The data were collected through semi-structured sociolinguistic interviews with 22 Hungarian Australians living in Sydney. The findings demonstrate that (a) first-generation Hungarians are highly creative language users and integrate a large number of English lexical items into their speech. Most lexical borrowings belong to the derivational blends with the highest proportion of the nominal group. Lexical borrowings from English are morphologically integrated with Hungarian-derivational suffixes and inflectional case markings. This research provides original empirical data to better understand the various inter-language lexical manifestations in Hungarian–English bilingual contexts. The study adds to the relatively small body of research on Hungarian–English bilingualism in diasporic context and contributes to understanding lexical borrowing from a contact linguistic perspective.

(here Hungarian) as a result of ongoing language contact (migration context) with another language (here English). These manifestations include morphological importation from the donor language (here English), or semantic shifts without importing foreign language material. Intra-language manifestations are not discussed here as they refer to linguistic phenomena which are not attributable to the impact of a foreign language.

Definitions and Terms
Researchers of bilingualism have approached the study of inter-lingual lexical phenomena from various diverse theoretical grounds and he differentiation between borrowing, code switching, and code-mixing has been a challenging task. Although multiword other language insertions can unambiguously be classified as code-switches (Poplack, 2012), the categorization of "Lone-Other-Language-Items" (hence, LOLIs) has long been the focus of debate in bilingualism studies, and several authors have argued for a less rigid distinction between the various terms. Poplack, Sankoff, and Miller (1988) were the first to address the problem of distinguishing borrowing from codeswitching of LOLIs based on a large amount of empirical data collected from bilingual French-English discourse (Poplack, 2012). Poplack et al. (1988) have found a large number of borrowings which showed the characteristics of code-switching, and a large number of code-switches which showed characteristics of borrowing. Based on their findings, they proposed the category of "nonce borrowing" (as something that is between borrowing and code-switching) and proposed four main criteria according to which loanword status can be established: (a) linguistic integration into recipient-language grammar, (b) diffusion across the community, (c) dictionary attestation, and (d) age of attestation (Poplack, 2012;Poplack et al., 1988). Appel and Muysken (2005) have argued for a distinction as introduced by Ferdinand de Saussure and stated that "borrowing involves the integration of two languages at the level of langue, and code mixing the integration at the level of parole" (p. 121).
A contrasting perspective is proposed by Myers-Scotton (2006) who has put forward convincing arguments as to why single word foreign language insertions should be considered as code switches. First, if they are pronounced according to the embedded language phonology (i.e., they are not integrated phonologically into the Matrix language), they are clearly code switches as "most established lexical borrowings are pronounced as words in the recipient language: that is, they show complete or at least partial integration into the recipient language" (Myers-Scotton, 2006, p. 256). Second, as embedded phrases (defined as more than one word) usually follow the word order of the embedded language, they can be code switches just like longer insertions. Third, embedded forms often appear in "bare forms" which means that they do not have the morphological inflections that would be necessary to integrate them into the matrix language. Such bare forms provide further evidence that they are code-switches rather than borrowings (Myers-Scotton, 2006). Matras (2009) sees borrowing and code-switching on two ends of the same continuum which is dynamic rather than strictly linear. Matras (2009) identified seven dimensions according to which the status of "foreign" language elements can be determined. These criteria include (a) the level of bilinguality (how proficient the speakers are), (a) the composition of the inserted element (how complex/simple the inserted structures are), (c) functionality (what conversational functions does the insertion serve, for example, stylistic choice vs. default expression), (d) unique referent (whether or not the lexical item is a core item or part of the para-lexicon, such as using the kinship terms on the target language, for example, Oma to a Syrian grandmother in Germany), (e) operationality (whether it involves core lexical or more automated grammatical items), (f) regularity (single vs. regular occurrence), and (g) structural integration (whether the inserted item is integrated phonologically, morphologically, or syntactically) (Matras 2009).
Without aiming to resolve these complex terminological and theoretical debates, for the purpose of this article, lexical borrowing is defined as the integration of foreign elements within the clause (intra-clausal) irrespective of the level of phonological integration. Full sentence or full clause length utterances were not counted as part of this study, but they were tagged for future analysis focused on conversational code-switching. This strategy was decided for two main reasons: First, the transcript was not accurate enough to determine the degree of phonological integration, and as some scholars have argued, phonological integration is not necessarily a reliable indicator of the borrowing status (see, for example, Kovács, 2005), and the phonological integration of borrowed items is "often gradual and highly variable" (Poplack, 2012, p. 644). Second, although the frequency of occurrence is often used as an indicator to determine the borrowing status (Myers-Scotton, 1993;Poplack, 2012;Poplack et al., 1988), the dataset collected was insufficient to speculate about the degree to which the various "nonce switches" (Myers-Scotton, 2006, p. 251) would have been established more broadly in the community.

Categorizing Lexical Borrowings
Working from the theoretical ground of contact linguistics, Winford (2003) categorized lexical contact phenomena into two main groups: (a) Lexical Borrowings and (b) Native Creations. This distinction is made on the basis of whether the lexical items in question are modeled on the donor language or not. Lexical Borrowings, as Winford states, "involve imitation of some aspect of the donor language" (p. 43), whereas creations "are entirely native and have no counterpart in the donor language" (p. 43). Winford further subdivided lexical borrowings into two groups: loanwords "in which all or part of the morphemic composition of the loan derives from the external source language" (p. 43) and loanshifts (loan meanings) where the morphemic composition is entirely native, but the meaning is at least partially derived from the donor language. Native creations include purely native creations which means that native words (in the current context, Hungarian) are used to describe foreign concepts, such as vasárnapi iskola [Sunday School] (Forintos, 2005(Forintos, , 2008; hybrid creations which are blends of native and foreign material to express foreign concepts (e.g., granny flat szállás-granny flat accommodation; Forintos, 2005Forintos, , 2008; and native creations using only foreign morphemes (no such examples have been reported in the Hungarian-Australian community). Forintos (2005Forintos ( , 2008) expanded Winford's categorization by adding a separate group, intra-lingual features, which included items where the impact of English was not detectable or where the change of usage or meaning was not modeled on the donor language. These items are discussed elsewhere as interference phenomena (Kontra, 1990) or neologisms (Kovács, 2005). This study draws in these theoretical and ontological considerations and follows the contact linguistics perspective to explore nuanced lexical manifestations in the Hungarian spoken in Sydney.
Another fundamental research agenda in language contact studies is to investigate the motivations behind borrowing. As bilingual speakers constantly make choices in terms of which part of their linguistic repertoire to activate, there are three main motivations behind borrowing: filling "gaps," prestige, and cognitive factors related to language processing (Matras, 2009). In terms of gap filling, the most frequently occurring fillers are related to cultural loans, where speakers do not have the translation equivalent term or they feel that the original term integrated from the other language expresses the phenomenon more unambiguously or authentically. Prestige has long been established in sociolinguistics 1 as a strong motivational factor in making linguistic choices which conform to the prestige variety and the usage associated with the dominant cultural group. 1 In the context of immigrant communities, borrowing from the language of the host society can be motivated by a broader acculturation strategy of fitting in and exiting the immigrant status. Third, psycholinguistic research on language processing in bilinguals has shown that it requires extra effort for bilinguals to keep their languages apart, as both languages are "on" all the time (Bialystok, 1999(Bialystok, , 2007, and more complex linguistic structures require more processing. This, however, more often relates to borrowing due to limited acquisition of L2. In the context of firstgeneration immigrants, this factor can be related to L1 attrition where the heritage language competence in some linguistic domains is becoming weaker and speakers are experiencing some form of linguistic insecurity. Finally, borrowability is a concept that refers to the likelihood in integration of certain linguistic elements into another language. In terms of borrowability, most research has focused on determining which word classes are more likely to be borrowed and whether there are any universals that apply to inter-lingual borrowability. As Matras (2009) has pointed out, prior studies in the field have applied varied methods and approaches, and current research is somewhat inconclusive about the categorization of borrowed items. Nevertheless, there are some tendencies which seem to hold truth across a number of language contact situations, such as the greater likelihood of open-class items to be borrowed compared with closed-class items (Matras, 2009). Matras studied 27 languages in contact and found the following frequency-based hierarchy of borrowings: Nouns > conjunctions > verbs> discourse markers > adjectives > interjections > adverbs > other particles, adpositions > numerals > pro-nouns> derivational affixes > inflectional affixes (Matras, 2009, p. 157)

Hungarian-English Language Contact Studies in Diasporic Contexts
This project drew on prior research in Hungarian-English bilingual diaspora in the United States and in Australia. The studies conducted in the U.S. Hungarian community (South Bend, Detroit, and McKeesport) used varied approaches for the differentiation between code-switching and borrowing. Kontra (1990) and Fenyvesi (1995Fenyvesi ( -1996Fenyvesi ( , 2005) used a phonological differentiation: Foreign language (English) insertions pronounced according to the donor language (English) were counted as code switches, while those pronounced according to the recipient-language (Hungarian) phonology were considered to be borrowings. Bartha (1993) used the frequency of occurrence as the basis for differentiation. Fenyvesi (1995Fenyvesi ( -1996 conducted interviews with 20 speakers of Hungarian living in McKeesport, Pennsylvania. These participants were mainly second-generation immigrants (16) and the remaining four were 56-ers or later arrivals. The interviews were about the subjects' life histories and patterns of Hungarian language use. Fenyvesi found numerous examples of non-Standard Hungarian (SH) case marking which is characteristic of the American Hungarian (AH). Fenyvesi (1995Fenyvesi ( -1996 gave a detailed account of two main tendencies as non-standard/deviant case marking, namely, the simplification in the use of locative cases in their primary, directional meaning and a unique pattern of the use of locative case marking with place-names: namely, the use of both the in-cases and the on-cases. Fenyvesi also observed evidence of language attrition as some of the deviant case marking was attributable neither to dialectal variation in Hungarian nor to the influence of English. Both American English and Hungarian dialects were discounted as the source for case of substitution, for example, in the following sentence: "egy évet jártam a University of Pittsburgh-nél" (SH -re) ["I went to University of Pittsburgh for a year"] (Fenyvesi, 1995(Fenyvesi, -1996 Throughout his study of the Hungarian used in the diaspora in South Bend, Indiana, Kontra (1990) contrasts American Hungarian (AH) with Standard Hungarian (SH). His analysis was based on 9 weeks of fieldwork which he conducted in 1980 and 1981. His respondents included first-generation immigrants such as those who migrated before World War I (WWI; "Old Timers"), and those who left Hungary at the time of the 1956 revolution against the Russian communist regime. In addition, the sample included second-and third-generation Hungarians as well (altogether he has 80 hr recorded from 40 respondents). Kontra's fieldwork gathered useful evidence of phonological, morphological, and syntactical changes, as well as lexical integration from English to Hungarian. In terms of sociolinguistic variation, Kontra noted the replacement of the formal second person singular pronoun (Ön, Maga = You formal) with the informal variety (Te = You informal).
Kovács (2001) studied code-switching and language shift in the Hungarian and the Finnish diaspora of Australia. Kovács (2001) found that the majority of code-switching involved intra-sentential switches where a nominal element (noun or noun phrase) was inserted into the sentence. These nominal expressions were typically in the subject position and they did not require any morphological endings (they were in the nominal case). Kovács also found that second-generation speakers used code-switching more often. Kovács (2005) reported the language use and bilingualism features of 18 Australian Hungarians attending the scouts club in Sydney. Language use was measured in five different domains, including home, church, neighborhood, work/school, and leisure. The study found a significant decrease in the use of Hungarian from the first to the second generation. Kovács (2005) gave a detailed account of the linguistic features of Australian Hungarian including inter-and intra-language features on phonological, morphological, and syntactic levels. Forintos (2005Forintos ( , 2008 studied the language contact phenomena in a corpus which constituted advertisements found in 98 issues of the Australian published periodical "Hungarian Life" (Magyar Élet) and identified numerous lexical borrowings. As this research was based on published written texts, the current study complements it by looking at naturally occurring spoken discourse.

The Hungarian Community in Australia
When discussing contact-induced changes in the use of a heritage language, it is important to consider extralinguistic factors (Clyne, 2003;Matras, 2009), such as attitudes, demographics, and the institutional support vis-à-vis which have been emphasized in various ethnolinguistic vitality theories (Allard & Landry, 1986, 1994Bourhis, Barrette, & Keith, 2006;Giles, 2001). Hungarians occupy the middle ground in terms of their language maintenance and shift in Australia (Clyne, 2003;Hatoss, 2006Hatoss, , 2008. According to the 2011 Census, 19,092 Hungary-born people live in Australia today (Australian Bureau of Statistics [ABS], 2011; see www.abs. gov.au). Hungarian migration to Australia has a long history, with a few Hungary-born migrants registered at the time of the 1850s gold rush and some who arrived after 1848. The largest number of Hungarians, however, started to arrive between 1945 and 1947 as Displaced Persons (DPs) who fled Hungary as a result of WWII and among whom there was a significant number of Holocaust survivors. Another significant wave of migration was the result of the 1956 revolution which brought 14,000 Hungarian refugees to Australia.
The Hungarian diaspora is one of the oldest post-war diaspora in Australia (median age = 66), which means that the ratio of those above 65 is one of the highest among those who were born overseas. This demographic has a direct influence on intergenerational language maintenance, as Hungary-born residents are rapidly decreasing in number. According to Census 2011, in comparison with Census 2006, there was a fall of 5.3% in the Hungary-born population. In terms of ancestry, 87.9% of Hungary-born Hungarians considered themselves Hungarian, 2.8% as German, and 1.3% as Jewish. In terms of language use, 59.2% spoke Hungarian, 35.9% English, and 1.5% German in their homes. In terms of English skills, 87.8% speak English well or very well and 10.4% speak English not well or not at all. In terms of religion, 54.6% are Catholic, 7.1% are Jewish, and 5.6% belong to the Presbyterian and Reformed Churches (ABS, 2011).
Another demographic factor which impacts on ethnolinguistic vitality is the fact that the community is rather dispersed across various suburbs in Sydney. Nevertheless, Hungarians have strong community organizations and cultural groups including the NSW Hungarian Association, the Hungarian House (Magyar Ház) in Punchbowl, the South Hungarian Club (Délvidéki Magyar Szövetség) in Glendenning, and many others. The programs in the Hungarian House in Punchbowl are mainly targeted at the old generation (mainly those who migrated after the WWII and in 1956 or soon after); the South Hungarian Club has a broad range of programs for every generation such as folk concerts, folkdance, picnics, and balls. They also celebrate all Hungarian national as well as all religious holidays keeping with the Hungarian traditions. Hungarian religious masses are held in several locations in Sydney, including the all-faiths masses in Strathfield, at the Uniting Church. The ethnic journal titled "Hungarian Life" (Magyar Élet) is the only Hungarian national weekly newspaper in Australia and it is sold in several news agencies across Sydney. These objective ethnolinguistic vitality factors play an important role in ensuring that first-generation Hungarians have the opportunity to maintain their heritage language.

Data Collection
Data were collected through various Hungarian community organizations such as the Hungarian House in Punchbowl and the Délvidéki Magyarok Egyesülete [South Hungarian Club]. A flyer was placed in popular Hungarian venues to recruit potential participants. The strategy was to recruit participants from a range of different venues across the community and from various suburbs to ensure a diverse and impartial outcome. Another strategy to broaden the dataset was to engage some respondents who lived in age-care facility. By applying a snowball technique, 22 first-generation Hungarians were recruited. These participants had a relatively regular contact with the Hungarian community and some of them knew each other. The interviews were semistructured and encouraged respondents to speak freely around the interview themes. In addition, basic demographic data about place and year of birth, occupation, and self-rated language abilities in English and in Hungarian were collected as part of the face-to-face interview. These self-ratings of language abilities were elicited through a custom-designed assessment tool which was based on the standardized International Second Language Proficiency Ratings (ISLPR). The ISLPR is used broadly in Australia for assessing immigrants' English language proficiency in reading, writing, listening, and speaking. This self-assessment served the purpose of identifying aspects of language dominance. All participants nominated Hungarian to be their stronger language in speaking.
The interviews were approximately 1 to 1.5 hr long, and they were conducted in Hungarian with a bilingual facilitator who was born in Hungary. Participants were told, however, that they were free to switch to English or any other language if they wished to do so at any time during the interview. Therefore, sociolinguistically speaking, the base language (unmarked code) was Hungarian, and psycholinguistically, we argue that the stronger language (dominant; Appel & Muysken, 2005) was also Hungarian for all the participants. Participants were not given explicit information about the focus on inter-language phenomena in their speech. Instead, they were informed about the broader aims of the project such as to study language maintenance in the community and the effects of language maintenance in older age. Nevertheless, as the project was conducted by linguists for linguistic purposes, it needs to be recognized that the frequency and degree of integrating English material into the Hungarian matrix language could have been affected by the presence of the interviewer. This possible bias was also reported in the context of American Hungarian (see, for example, Kontra, 1990).

Data Analysis
All identified relevant tokens were entered into an Excel database for linguistic categorization and coding. The strategy was to tag items according to the word class categories, the speaker's identification code, and the type of lexical borrowing according to Winford's (2003) categories. Similar approach was also used by Forintos (2005). In the second phase of the analysis, those lexical features were identified which did not incorporate foreign material, but they were nevertheless deviant from the SH usage. For this purpure, the Hungarian National Corpus (HNC, http://corpus.nytud.hu/ mnsz/index_eng.html) compiled by the Hungarian Academy of Sciences was used. This corpus contains 187.6 million words, and it is divided into five sub-corpora according to regional variation and five sub-corpora according to genres. Although the discussion that follows includes some quantified data, the main approach taken here is a qualitative analysis.

Participants
Altogether 22 people (10 males and 12 females) participated resulting in a corpus of 18 separate interviews which yielded 106,094 words of transcribed material. Most interviews were conducted in participants' homes, while some (n = 6) were collected in the South Hungarian Club (DMSZ) and some (n = 3) in the St Elizabeth Aged care Facility. All participants nominated Hungarian as their mother tongue and the language that they learnt from birth. In terms of ethnicity, two of the participants were of Jewish background. In terms of age, the oldest was born in 1917 (age = 98) and the youngest was born in 1951 (age = 64). The average age at arrival was 16.3 years with the youngest being 14 and the oldest 40 at arrival in Australia. In terms of place of birth, participants came from a broad range of geographical areas in Hungary (Budapest = 4,Győr,Zalaegerszeg,Esztergom,Gyomaendrőd,Lenti,Debrecen,Dorog,Szárföld,Veszprém,Mór,Nyír,and Pásztó) and some came from parts of historical Hungary which are now part of Serbia: Bácsföldvár, Újvidék [Novi Sad] = 2, and Óbecse. In terms of the vintage of migration, three participants arrived in Australia in 1949, 11 in the 1950s, five in the 1960s, and one in 1970. Participants resided in various suburbs of Sydney including Castle Cove, Bardwell Valley, Illawong, Arncliffe, Bondi, Bexley, and Summer Hill. Interviews were conducted between February and May 2015. See Table 1 for the demographic data of participants.

Lexical Borrowings
Considering that the research collected spoken data through a sociolinguistic interview, lexical borrowings were categorized first in terms of the classic word class categories (nouns, verbs, adjectives, etc.), but discourse markers were also counted. Phrases which composed of more than one single lexical item or word (e.g., compounds, noun phrases, and adjectival phrases) were counted separately only if both elements of the constituent were borrowed from the other language (here from English). In addition, various idiomatic expressions and discourse markers were counted separately, following the method applied by Kovács (2001).
These discourse markers (n = 56) included interjections (e.g., Yeah!) and any other expressions which served the evaluative function in the discourse, such as "Oh God!" and so on. In addition to the "insertional code switching," the full database also included numerous examples of "alternational code-switching" (Matras, 2009, p. 101). These were full clauses which were inter-sentential or supra-sentential codeswitches and were excluded from the scope of this article. See Table 2 for the quantitative summary of borrowing by parts of speech. The findings show the greatest number of borrowing in the nominal group, such as single nouns (168) and noun phrases (56). These are followed by discourse markers (56) and adjectives (48) and verbs (39). Proper nouns (44) were counted separately, and these have no significance in terms of borrowability.. Adverbs (18) and idioms (4) were the smallest group of word class categories, while there were no numerals borrowed. This finding is consistent with prior studies in Hungarian-English language contact (Bartha, 1993;Kontra, 2001;Kovács, 2005).
All items which did not fit the straightforward categories of parts of speech were classified under the generic heading of discourse markers. This means that interjections and discourse particles were not differentiated but they were all tagged under this code.
By categorizing the L2 (English) items according to parts of speech, the analytical strategy was to account for the distribution of borrowing in syntactic as well as in discursive terms. Prominent semantic fields were also identified and are discussed below. Lexical borrowings were tagged and categorized following Winford (2003): (a) Loanwords (pure loans and derivational loans) and (b) Loanshifts (semantic loans/extensions and loan translations [calque]). Altogether there were 153 pure loans which fit the description of "total morphemic integration of single or compound words, various degrees of phonemic substitution" (Winford, 2003, p. 45). There were 223 examples in the corpus which fit the category of loanblends (combination of native and imported morphemes) with the majority being derivational blends (imported stem + native affix). As derivational blends were numerous, it is not possible to list them in this article. Some of the derivational blends are listed in the appendix.
Adjectival derivatives were relatively rare and included four examples: high school-os időbe [during high school years], subtitle-ös filmet [SH feliratos, film with subtitles], excited-ok voltunk [we were excited], and upset-ok voltunk [we were upset]. Those items which included an imported stem and a native stem were categories as compound blends (Winford, 2003). These examples were rare, similar to Forintos's written

Semantic Fields of Derivational Blends
The dataset showed a large number of inserted items which were proper nouns such as names of institutions or localities, and these were counted separately following Kovács (2001). The frequent use of L2 institutional terminology in L1 in migration contexts has been documented in the literature (Matras, 2009). The reason for such insertions lies in the fact that "the original name evokes associations with the original setting and allows the speaker to import the image of that setting directly into the context of the ongoing conversation" (Matras, 2009, p. 107). Most of the proper nouns included place-names (Sydney, Melbourne, Adelaide, Bonegilla, Hyde Park, etc.), names of companies (e.g., Myers, David Jones, etc.), and names of educational institutions: 1. ő jár a University of New South Wales-re he/she go University of New South Wales-SUBL "he goes to the University of New South Wales" 2. a Charles Sturt University-n körösztül Charles Sturt University-SUP-through "through Charles Sturt University" 3. a Sydney Technical College-ba Sydney Technical College-ILL "at the Sydney Technical College" One reoccurring feature of these items was that some of the inflectional endings expressing the locative were nonstandard, such as replacing the SH [-ban/-ben] expressing a static locative function expressed in the Inessive case marking, with [-ba/-be], the non-standard derivational forms which are normally used to express the directional locative (Illative case). This is, however, a typical characteristic of spoken Hungarian in Hungary and is not limited to the diaspora in question. Similar findings were reported by Kovács (2005). The examples collected in this corpus belonging to the semantic field of place-names included the following: Another semantic group of lexical borrowing was the integration of items referring to Australian cultural phenomena, particularly those relating to their migration history and their first experiences. This is consistent with the gap-filling or cultural loan theory (Matras, 2009). These lexical items typically had no Hungarian equivalent; therefore, the borrowed items filled a lexical gap. Such items referred to Australian national holidays, such as "ANZAC Day-en érkeztem meg" [I arrived on ANZAC Day]. The lexical borrowing "wog" was used to reminisce about the identity that they were assigned by Australians. "Wog" is an Australian slang which has a derogatory meaning and was popular after WWII to refer to immigrants, particularly those with olive skin such as Southern Italians:

"Az éppen olyan wog mint én vagyok"
just like wog I-BE-1SG "He is just as wog as I am" 5. "Nem akarta a nyakláncot, mert azt csak wog boy hordja" no want necklace because that only wog boy wear-3SG "He didn't want the necklace, cause only a wog boy would wear that" The term New Australian was another commonly used expression to describe post-war immigrants and participants reminisced about this during the interviews: 6. "bloody new Australian-nak hívták őket" bloody new Australian-DAT call-3PL PAST them "They were called bloody New Australians" Another commonly occurring expression was "DP" (pronounced as "dipi") which was used to refer to post-war refugees. "DP" was used as an adjective in the following: 7. "még talán az is meg van, a passport, a dipi passport" still maybe that BE-3SG, the passport, the dipi passport "I may still have the passport, the DP passport" Another dominant semantic field included expressions referring to housing and places of residence or significant places or institutions. These topics naturally occurred as participants were asked to talk about their memories of how they settled, where they lived, and about their memories of their own and their children's schooling. Some of these were specific referents which were either difficult to translate to Hungarian, or the translation equivalents were not triggering the same meaning and therefore would have compromised the clarity and the authenticity of the phenomena. These examples are provided with English spelling below (see, for example, day-care center), as their pronunciation was closer to the English original: 8. ilyen day-care centre-be kicsikkel such day-care-centre-ILL small ones-COMitative "In sort of a day-care centre with little children" 9. egy ilyen boarding house-ba mentünk lakni a such boarding house-ILL go-1PL-PAST live "we went to live in sort of a boarding house" 10. utána meg a nursing home-ba tudok menni then nursing home-ILL can-1SG go-INF "then I can go to the nursing home" 11. ott volt ilyen housing commission házunk there be-PAST housing commission house-POSS-PL "there we had a housing commission house" 12. igen a granny flat-be yes granny flat-IN "yes, in the granny flat" There were a large number of derivational blends which described language behavior. Some of these referred to participants' level of English: 13. olyan level-ön voltak such level-SUP BE-3PL-PAST "They were on that level" Although the word "accent" is an established loanword in SH (akcentus), participants frequently used the English borrowed version "accent" (pronounced as "akszent") with Hungarian-derivational suffixes, such as the following: In this semantic group, the verbal stem /spell-/ occurred by far most frequently. Kovács's study also found this to be a frequently used stem for borrowing. The verb "spell" was integrated morphologically into Hungarian using the derivational verbal suffix /-l/. Kovács (2005) found that /szpellöl/, /szpellol/, and /leszpellingel/ ("spell") were found frequently in the Australian Hungarians' speech. She also noted that there was a variation in terms of word formation and this was also evident in the current corpus. For example, "spelling" was integrated in a variety of forms including /szpellelés/, /szpellolás/, /szpellölés/, /szpelling/, /szpellingelés/. In terms of the verbal form, /to spell/ was integrated as /szpellingel/, /szpellingol/, and

Verbs
The verb forms were most frequently integrated morphologically into Hungarian using the derivational verbal suffix -l. This was consistent with previous studies conducted in other Hungarian diaspora contexts, such as the Burgenland Hungarians in Austria (Gal, 1989), Hungarians in the United States (Fenyvesi, 1995), and in the speech of Australian Hungarians (Kovács, 2005). The examples below are derivational blends where the English root is "Hungarianised" in the way that the Hungarian-derivational morpheme "l" is attached to the end of the word to "make it" a Hungarian verb, but the suffix is preceded by a linking vowel. 49. átszvicsölök (preverb át "through" switch-ö-l-1SG) 50. lecsekköl (preverb le-"down" check-ö-l-3SG) 51. letájpolta (preverb le-"down" type-o-l-3SG-PAST) There were some examples where the -gat/-get derivational suffix was added to the derivational suffix -l to express a repeated action. In the first example below, this derivational blend was created from the English stem "pat" (e.g., to pat on the shoulder). In the second example, the English verb stem "force" was added the derivational suffix -l (ol), followed by -gat and completed by the first person plural inflectional marking, which in this present case is a non-SH usage (-suk >< SH -juk), and is typical of informal or dialectal speech.
52. akkor mindenki sírtak, és jöttek, és pet-o-l-gat-tak utána, "All of them cried and came to pat me on the shoulder afterwards" 53. nem forsz-o-l-gas-suk "we are not forcing" The English present continuous (progressive) tense also appeared in some of the blended verbs: 54. édesanyám living-e-l "My mother is living" 55. ríszőrcsing-ö-l-i "is researching" 56. májnding-o-l-o-m "I am minding" There were a small number of examples of a derivational blend which used the Hungarian -va/-ve derivational suffix to form a past participle of a passive construction, such as, 57. üzletileg involvolva vagyunk (involve-o-l-va) "we are involved through business" 58. lecsekkölve (le-csekk-ö-l-ve) "checked out" Adverbial derivational blends were relatively rare and included the use of the Hungarian adverbial derivational suffix -leg (such as quality-leg), -n (such as funny-n), and -ul/ ül (such as tough-ul)

Semantic Loans
As stated in the introduction, loanshifts are those borrowings which only contain native material; that is, the native Hungarian words are used, but their meaning is extended and this extension or change is traceable back to the donor language, in the context of this article, English. Loanshifts were divided into semantic extensions (shifts in the semantics of the native word) and calques (loan translations). Previous studies (Fenyvesi, 1995;Forintos, 2005Forintos, , 2008Forintos, , 2012Forintos, , 2015 have found a large number of semantic extensions (or semantic loans) in their corpora collected from the Australian Hungarian and AH diaspora, respectively. In fact, semantic extensions were the second largest group of lexical borrowings.
In the excerpt below, the respondent was asked if his children used Hungarian properly. In his response, he used the Hungarian word "korrektul," which is otherwise an established loanword in SH and was used with the meaning of "correctly." This is clearly a result of the interference with English. The SH adverb "korrektul" means "fairly" or "in a correct fair manner" and cannot be used with the meaning of "correctly," as in "grammatically correct": Further examples of semantic extensions are listed below and show that the Hungarian verb "dolgozni" [work] was used instead of the SH "működni" [function/operate/work] to describe that something "operates" or "works out": 59. kórház már nem dolgozik ott ENG: "The hospital does not operate there anymore." SH: "nem működik" OR "nem üzemel" 60. az nem fog dolgozni ENG: "this will not work" SH: "az nem fog működni"

Loan Translations (Calques)
A specific sub-group of semantic loans included calque (loan translations) where there was evidence of the combination of native (Hungarian) morphemes in imitation of the English pattern. In other words, these items were literal translations from English into Hungarian to create new non-SH expressions. For example, the literal translation of "birthday card" was used in Hungarian (születésnapi kártyá-tbirthday card-ACC), instead of the SH "üdvözlőlap." The Hungarian word "kártya" is used in compounds such as bank card and playing card, but not for a greeting card. In another example, the speaker translated the phrase "one-bedroom apartment" into Hungarian literally ("egyszoba lakás-t-one bedroom apartment-ACC"), instead of using the SH adjectival phrase ("egyszobás lakás-t-one-bedroom apartment-ACC"). The lack of the adjectival derivational suffix -s shows the influence of English adjectival derivational rules where the noun can function as an adjective without any suffixation. The English phrasal verb "look after" has the SH equivalent of "gondoskodik + Indirect object-DELative." In the example below, the speaker literally translated from English to arrive at the non-standard use of "utánanéz." This verb exists in Hungarian with the meaning of "to look into something" or "find out," but does not refer to "minding" or "taking care of." 61. apám négyszer nősült, és mindig kaptam egy-egy új személyt, aki utánam nézett "my father married 4 times and I always got a new person who looked after me" SH: "gondoskodott rólam" Below is a list of further examples for semantic loans involving literal translations (calque): 62. Akkoriban szépen letettük a fejünket "in those days we put our heads down" SH: "lehajtott fejjel/fővel jártunk" 63. jól fölszedte a magyart "picked up Hungarian well" SH: "jól elsajátította a magyart" OR "jól ráragadt a magyar" 64. került a speciális iskolába "went to a special school" SH: "kisegítő iskolába került" 65. kidolgoztam magamban "worked it out (by) myself" SH: "magamtól rájöttem" 66. szerencsénk van ezen az oldalon "we are lucky on the one hand (in this regard)" SH: "ilyen szempontból" 67. éjszakai technikumba járt "attended evening college/classes" SH: "esti iskolába" The verb "dolgozni" also showed non-standard usage when appearing with the verb "kimegy" ["go out"]. In SH, the preverb el-[away] is used to refer to going to work, rather than the preverb "ki" [out]. This is another example of a straight translation from English into Hungarian.
68. anyuka nem ment ki dolgozni, "mum didn't go out to work" SH: nem ment el dolgozni/nem járt el dolgozni In the next calque, the Hungarian verb "dolgozni" showed non-standard usage in terms of its adverbial collocation and this is most likely due to the word for word translation of the English expression "work hard" into Hungarian. In the following example, the Hungarian adverb "nehezen" ["hard," "with hardship"] was used as a collocate instead of the SH "keményen" ["hard"]. The collocate "nehezen" [hard] is used in SH in examples such as "nehezen ment a munka" [hard go-3SG PAST work-NOM] and "nehezen bírták a munkát" [hard bear-3PL PAST work-ACC] (HNC).

Dolgoztunk nehezen
"we worked hard" SH: "nehéz munkát végeztünk" OR "keményen dolgoztunk" In the following example, the effect of the English preposition "through" is shown in the use of the Hungarian verbal prefix "keresztül." In SH, the verbal prefix "át" is used with the verb "olvas" to mean "read through something." 0. keresztülolvasom még egyszer "I will read it through again" SH: "átolvasom" The preverbal construction "kimenni" has the literal meaning of "go out," such as go out of the house in SH. However, in the example below, the meaning was extended under the influence of English to refer to the children who have left home: 71. gyerekek kimentek a házból "the children left home" SH: "elköltöztek otthonról" Additional similar examples are listed below: 72. amelyik éppen könnyebben a fejedbe jön "whichever comes to your mind/head" SH: "eszedbe jut" 73. vissza kell tölteni egy formot "fill in a form back over again" SH: "ujra ki kell tölteni egy formanyomtatványt" 74. az angolom kezd szenvedni "my English is starting to suffer" SH: "romlik az angol tudásom" 75. Elmentünk mozikba (plural) "we went to the movies" SH: "elmentünk moziba" (singular) 76. akkor engem nagyon szépen kezeltek az olaszok "Then the Italians treated me very nicely" SH: "Jól bántak velem az olaszok"

Conclusion
This research aimed to contribute to the literature on Hungarian-English language contact in diasporic contexts. The study has identified numerous inter-lingual lexical manifestations in the speech of first-generation Hungarian immigrants living in Sydney. These manifestations were grouped according to Winford's (2003) classification of lexical borrowing phenomena. From the findings, it can be concluded that first-generation Hungarian Australians use a wide range of lexical borrowing in their speech and these borrowings are morphologically and syntactically integrated into their Hungarians. Their speech, as discussed in this article, was impacted by English in two ways: (a) morphological importation of English material and (b) semantic shift, where the meaning of Hungarian words shifted due to the exposure to English. These findings demonstrate that first-generation Hungarian-Australian bilinguals use their linguistic creativity to express themselves in bilingual sociolinguistic contexts, such as the sociolinguistic interviews, and make linguistic choices to draw on their full linguistic repertoire to best fulfill their communicative intensions. On the other hand, non-standard usage points towards language attrition which needs further exploration. The word class categories confirm prior research that the nominal group is borrowed most frequently. The semantic fields to which the various borrowings belong not only reflect the topics of the interviews but also the prominent themes which are memorable for immigrants (such as housing, schooling, and employment). Some other semantic fields reflected the bicultural and bilingual congruity where Australian cultural events and significant places were named in English, as opposed to seeking translation equivalents in Hungarian. The exact nature of the motivation behind these choices cannot be established based on this dataset. However, the examples are reflective of duality of their cultural and linguistic everyday practices. The lexical borrowings identified in this corpus are indicative of the types of inter-language manifestations that occur in Hungarian-English diasporic contexts, especially in Australia. The morphological integration of lexical borrowings contributes to our understanding of the general theory of grammar from a Hungarian-English contact linguistics point of view (Matras, 2009). However, based on the current dataset, the contact phenomena identified cannot be categorized in terms of their level of establishment and frequency of use. A larger sample size is required to examine the extent to which LOLI can be considered "nonce borrowings" or more established loan words in "Aussie Hungarian." As the data in this study were restricted to first-generation speakers, further research is necessary to examine the inter-lingual lexical manifestations in the speech of bilingual second-generation speakers which could provide an insight into some of the common and distinctive trends across the two generations.
In terms of the theoretical implications of language contact, this study has not aimed to refute current theories of lexical borrowing and has not aimed to resolve the long-standing issue of differentiating between lexical borrowing and code-switching. Although the distinction between these two terms and approaches is important for clarity and congruency in language contact studies, it is important to explore the various inter-lingual phenomena (whether classified as code-switches or borrowings) from the functional sociolinguistic perspective and continue to theorize the diverse range of factors that impact language choice in bilingual and multilingual diasporic contexts. The inter-lingual data collected through this research provides a preliminary dataset for a broader investigation of language use in diasporic contexts of Hungarian-English bilingualism.